Context: Ectopic pregnancy (EP) is an important cause of maternal deaths in early pregnancy, as most fatal cases resulted from delayed diagnosis and inappropriate investigation.

Objective: We evaluated whether the measurement of activin-A may be useful in the diagnosis of EP in women with unknown pregnancy location (UPL).

Design: Open observational study.

Setting: Tertiary referral center for obstetric care.

Patients: UPL women (n=536) with complaints of bleeding, pain or cramping.

Interventions: clinical examination; transvaginal ultrasound scan; hCG, progesterone and activin-A measurements; laparoscopy; uterine curettage; histological examination.

Main Outcome measures: Pregnancy outcomes; evaluation of sensitivity, specificity, and predictive values of hCG, progesterone and activin-A as diagnostic tests for the detection of EP.

Results: Pregnancy outcomes included 155 (28.9%) viable intrauterine pregnancies (IUP), 305 (56.9%) first-trimester spontaneous abortion (SAB), and 76 (14.2%) ectopic pregnancies (EP). SAB had lowest (P<0.0001) hCG and progesterone concentrations, significantly lower than EP (P<0.001) and IUP (P<0.001). In EP levels were significantly (P<0.001) lower than in IUP. On the contrary, activin-A levels were lowest (P<0.0001) in EP, significantly lower than in SAB (P<0.001) and IUP (P<0.001). IUP had significantly (P<0.001) lower activin-A levels than SAB. When evaluated by the Receiver Operating Curve (ROC) curve analysis, activin-A at the cut-off of 0.37 ng/mL combined a sensitivity and a specificity of 100% and 99.6%, respectively, for prediction of EP. When activin-A concentrations were below the cut-off, the positive predictive value for EP was 97.43%, and 0% for concentrations higher than 0.37 ng/mL.

Conclusions: Activin-A measurement may identify patients at risk of EP with a high sensibility and specificity.